论文部分内容阅读
目的探讨ARIMA乘积季节模型在HFMD月发病率预测中的应用价值,为手足口病防控工作提供依据。方法以2009年1月~2015年12月全国HFMD月发病率数据为基础,使用Eviews8.0建立ARIMA乘积季节模型,采用2016年1月~6月实际数据验证,以RMSE、MAE、Theil IC、BP、VP、CP、MRE评估模型预测结果。同法预测2016年7月~12月全国HFMD月发病率。结果 2009年1月~2016年6月HFMD月发病率最优模型是ARIMA(1,0,1)×(0,1,1)_(12),模型口径为:(1-B~(12))(1-0.574B)xt=(1-0.441B)(1+0.919B~(12))ε_t,RMSE=4.15,MAE=2.59,Theil IC为0.143,BP=0.045,VP=0.225,CP=0.730,2016年7月~12月发病率(1/10万)预测值分别为24.70、14.96、15.00、14.12、13.05、11.22。结论 ARIMA(1,0,1)×(0,1,1)_(12)模型可较好地拟合全国HFMD月发病率的演变趋势,可用于HFMD发病率的短期预测。
Objective To investigate the application value of ARIMA product seasonal model in the prediction of monthly incidence of HFMD and provide basis for the prevention and control of HFMD. Methods Based on the data of monthly incidence of HFMD from January 2009 to December 2015 in China, ARIMA product seasonal models were established using Eviews8.0. The data were validated from January to June 2016 with RMSE, MAE, Theil IC, BP, VP, CP, MRE assessment model prediction results. The same law predicted July 2016 ~ December 2016 national HFMD monthly incidence. Results The best monthly incidence rate of HFMD from January 2009 to June 2016 was ARIMA (1,0,1) × (0,1,1) _ (12). The caliber of model was (1-B ~ (12) RMSE = 4.15, MAE = 2.59, Theil IC was 0.143, BP = 0.045, VP = 0.225, CP = 1-0.441B) = 0. 730, and the predictive value of incidence (1/100) in July-December 2016 was 24.70, 14.96, 15.00, 14.12, 13.05, 11.22 respectively. Conclusion The ARIMA (1,0,1) × (0,1,1) _ (12) model can well fit the trend of HFMD monthly incidence in China and can be used for short-term prediction of the incidence of HFMD.